Using Data Science Power for Common Good

When the UMass Amherst campus entered its three-month-long stretch of summer tranquility, thirteen UMass students joined the Manning College of Information and Computer Sciences' Center for Data Science (CDS) to use their data science, software engineering, and technical writing skills to serve the common good, tackling seven data science projects focused on environmental impact, social justice, and technological ethics issues. 

Data Science for the Common Good (DS4CG) is a 12-week summer internship program in which undergraduate, master's, and doctoral students work with nonprofit and public sector organizations on data science projects. These talented students are led by CDS staff and work closely with partner organizations on data visualization, ingestion, modeling, application development, and other analytic tasks. 

The 2023 DS4CG participants attended seminars related to ethics, diversity, and data science best practices led by UMass faculty. These seminars included a dive into the rich history of W.E.B. Du Bois' data visualizations by Adam Holmes, the assistant director of the W.E.B. Du Bois Center, as well as a data science ethics training by professor Michelle Trim. 

Each team of students focused on helping an organization with a problem that benefits the common good. For example, in the aftermath of a disaster, the Red Cross needs an accurate count of damaged buildings to provide relief to the affected area. One DS4CG team joined a partnership that began the previous summer to continue work on a labeling tool for Red Cross Netherlands. In collaboration with the Computer Vision Lab, DS4CG participants Aditya Salveru and Debanjan Mondal successfully integrated the DISCount algorithm into a satellite imagery labeling tool that can take pre- and post-disaster images and automatically estimate counts and confidence intervals of damaged buildings. This lightweight model runs in a browser and will save the Red Cross significant time and resources in their mission to deliver rapid aid to those in need. 

This partnership with the Red Cross is just one example of how DS4CG participants take data science out of the lab and into the community. 2023 DS4CG teams also worked on improving energy efficiency in buildings through predictive modeling, tracking bird migration patterns in weather data with machine learning, detecting extreme speech in YouTube videos, extracting bylines from news articles in multiple languages, generating metrics for the center's Unity cluster, and tracking #StopAsianHate on the social media platform X.

To meet this year's DS4CG participants and learn about this summer's projects, stop in to the DS4CG Fall Poster Session on Wednesday, September 14, between 11:00 a.m. and 1:00 p.m. in LGRC room A112.